Fluid Flow Mass Transport for Generative Networks
This addresses training challenges in generative models for AI researchers, but appears incremental as it modifies existing methods.
The paper tackles the slow convergence and difficulty of training Generative Adversarial Networks by proposing a new formulation based on strict minimization, viewing it as a matching problem rather than adversarial, which allows for quick convergence and meaningful metrics.
Generative Adversarial Networks have been shown to be powerful in generating content. To this end, they have been studied intensively in the last few years. Nonetheless, training these networks requires solving a saddle point problem that is difficult to solve and slowly converging. Motivated from techniques in the registration of point clouds and by the fluid flow formulation of mass transport, we investigate a new formulation that is based on strict minimization, without the need for the maximization. The formulation views the problem as a matching problem rather than an adversarial one and thus allows us to quickly converge and obtain meaningful metrics in the optimization path.